Research associated with cross-sectional research, time-series research, and case studies involves collecting data that can test whether the actual behavior of individuals and organizations is consistent with the hypotheses in the theory. Much research in social science and management develop and test theory in this manner. However, scholars who believe that the current practices occurring in companies are not desirable or optimal should engage in "action research." In developing two new management accounting approaches-activity-based costing and the balanced scorecard, Kaplan and his colleagues used a particular form of action research, refereed to as "innovation action research." Innovation action research occurs when scholars are actively engaged in helping organizations to implement new ideas. Through this process, Kaplan and his colleagues identify a new concept and continue to apply and improve the concept through publication, teaching, and active intervention in companies.
In this paper, Kaplan draws on his personal experiences during the past 15 years to propose a preliminary theory for innovation action research. The experiences helped to create new management accounting theories and practices that are being applied today in organizations throughout the world.
Professor William Cooper argues that research intended to solve a specific problem in practice should be considered "basic" if it can be generalized beyond the initial application (1) to applications that are different from the one initially addressed and directed and (2) to alter the state of scientific knowledge as well as practice.
Argyris (1997) has formulated several features of scholarly consulting done by him and other psychologists:
1) Propositions (i.e., proposed solutions to organizational problems) that are valid and actionable, that are generalizable and applicable in the individual case. (i.e., the proposed concept is applicable to a broad range of organizations and can be customized and made actionable for targeted, specific organizations.)
2) Propositions that can be produced by scholars or practitioners under real-time, everyday-life conditions. (i.e., the concept is applied in actual situations, not in a research or laboratory setting. It also requires that the concept be stated simply enough so that it can be implemented in organizations.)
3) The effective implementation of these propositions leads to results that are consistent with what is intended. (i.e., expected unintended or unstated side-effects are minimal.)
4) The claim of effectiveness is testable by methods that meet the most robust standards of disconfirmation. (i.e., being able to effectively implement new ideas in organizations.)
Next, Kaplan discusses how the innovation action research followed for activity-based costing and the balanced scorecard satisfies the criteria proposed by Cooper and Argyris.
INNOVATION ACTION RESEARCH CYCLE
The illustration below shows the various stages of the innovation action research cycle.
1st Loop Around the Innovation Action Research Cycle
Step 1: Observe and Document Innovative Practice
The innovation action research cycle starts by identifying plausible solutions to the documented shortcomings in existing practice. For activity-based costing and the balanced scorecard, Kaplan and others had documented the limitations with existing practice in cost and performance measurement through articles and books. In addition, they were teaching and speaking about these limitations with M.B.A. students and executives. This communication process led directly to the initial solutions to both problems.
The ideas for both activity-based costing and the balanced scorecard innovations arose in practice. The visibility Kaplan and Robin Cooper had from writing, speaking about and being members of professional associations (CAM-I, Manufacturing Studies Board) concerned with improving management practice provided them with several opportunities to identify the innovating practitioners and organizations (such as Scovill Corporation, John Deere & Co., Siemens, and Hewlett-Packard). At this stage of the innovation action research cycle, the identified innovation is at an early, somewhat crude form. It may not be generalizable and applicable to many other organizations without additional refinement.
Step 2: Teach and Speak About the Innovation
Cooper and Kaplan began to teach the new cases about early activity-based costing initiatives in their M.B.A. courses and executive programs. To teach the cases, they needed to prepare teaching strategies and teaching notes. The preparation and teaching processes motivated them to understand the underlying phenomena in a deeper, more systematic and more conceptual way. M.B.A.s and, particularly, executives come from highly diverse backgrounds-many different manufacturing and service industries, domestic and international, for profit and not-for-profit, general management, as well as from diverse functional backgrounds (such as general management, finance, marketing, operations, technology and staff). Taking a new idea to these experienced individuals, especially in an interactive teaching environment, provides an early test of the validity and generalizability of the new idea. They also used the experiences and their emerging conceptual framework from these cases to give talks at academic and practitioner conferences.
For the balanced scorecard, even more initial learning came from testing the ideas directly with a set of companies that participated in a yearlong project on performance measurement with the Nolan, Norton & Co. The project attracted senior financial and planning executives from a dozen companies who met on a bi-monthly basis throughout 1990. Analog's corporate scorecard captured the interest of the participants. Throughout the year, they experimented with it in their organizations and reported on the results. The concept proved successful in many of the pilot sites and turned out to be the prime output from the year-long research project. In the process, the original corporate scorecard, which focused mostly on operational improvements (on lead times, delivery performance, manufacturing quality and cycle times) had become transformed into a much more strategic organizational performance measurement system, characterized by four identifiable perspectives (financial, customer, internal business process and innovation and growth).
Thus, at the end of the second step of innovation action research, the initial theory or concept has become codified, generalized and shown to be applicable to a much larger audience than the originating set of companies.
Step 3: Write Journal Articles and Books
The innovation action research cycle continues by exposing the ideas to an even wider audience. Cooper and Kaplan did this for activity-based costing through papers aimed primarily at a practitioner audience. The articles appeared in Management Accounting, Journal of Cost Management and Harvard Business Review (Cooper and Kaplan 1988a; Cooper 1988a, 1988b, 1989a, 1989b; Kaplan 1988; Cooper and Kaplan 1988b). For the balanced scorecard, David Norton (who had served as the Nolan, Norton & Co. project leader and facilitator) and Kaplan published an article (Kaplan & Norton 1992) to summarize their findings. Their explicit goal in writing these articles was to generate some excitement, enthusiasm and debate about the new ideas among a broad management audience.
Step 4: Implement the Concept in New Organizations
The active involvement of the researcher with new implementations serves several critical functions:
1) Validates that the new concept is implementable beyond the initial set of companies with whom the concept was developed. Also, tests whether the concept can create value for these new organizations.
The researcher, following the innovation action research cycle, has accumulated considerable knowledge about how the new concept was developed and implemented in the organizations where the idea first emerged (as documented during the initial set of teaching cases). The action researcher has also accumulated considerable skills in teaching, theory creation and explication of the concept. Therefore, perhaps more than any other person, the action researcher may be in the best position to advise on the implementation of the idea in a new setting.
Management scholars cannot test their ideas in laboratories. Laboratories cannot simulate the complex settings, relationships and structures where the most important of new management ideas must be implemented. Companies and organizations are the laboratories for management scholars.
In summary, scholars engaged in innovation action research play an active role in implementing their ideas in actual organizations. The concepts must promise sufficient benefits (i.e., they represent a solution to a real problem) and be articulated clearly enough that organizations are willing to commit their own resources to an implementation experience.
2) Provides learning opportunities to advance knowledge about the concept.
The early formulations of a new concept are not complete. They represent what has been learned from the initial set of implementations.
Cooper, after systematically studying the variety of cost drivers in the population of organizations which had implemented activity-based costing, developed a powerful taxonomy for classifying the different types of cost drivers used in activity-based costing models. This taxonomy was a major breakthrough in the theory of activity-based costing.
For the balanced scorecard, senior managers in two companies took the original concept and applied it in their companies in innovative ways, well beyond what Norton and Kaplan had described in their initial article (Kaplan & Norton 1992). These experiences revealed how the balanced scorecard could become much more than a measurement system. It could be used as an organization's central management system. Rockwater used the scorecard to translate its mission and strategic objectives into operational measures that employees could follow to structure new customer relationships with targeted market segments.
3) Creates knowledge about the implementation process for the new concepts.
The researcher, by being actively involved in the implementation process, also learns a great deal about the conditions leading to either successful or unsuccessful implementations of the new idea. Especially in the early stages of concept development, implementations can fail for several reasons. They can fail because the concept has not been sufficiently well developed or is, in fact, invalid or inappropriate in a context different from where it had been initially developed and applied. The concept can also fail, however, because managers or the advocates were not skilled in using the theory underlying the concept to guide specific managerial actions, i.e., how bad things happen to good ideas. Such failures often occur because of personal or organizational resistance, or lack of knowledge about the concept. In such cases, the implementation fails, not because of inherent limitations of the concept, but because of a poor implementation in the specific setting (Shields and Young 1989; Argyris and Kaplan 1994).
As Cooper and Kaplan worked with several firms, observing both their successes and failures, they learned a great deal about the problems that arise during activity-based costing implementations. For the balanced scorecard, Norton's and Kaplan’s experiences at several organizations gave them insights about the conditions for success, as well as failure, with using the balanced scorecard in organizations.
2nd Loop Around the Innovation Action Research Cycle
Step 1: Observe and Document Innovative Practice:
This comes naturally from the scholar's involvement with the newest implementations and enables these experiences to be captured in a new set of cases and teaching notes.
Step 2: Teach and Speak About the Innovation:
The new cases are taught to M.B.A.s and executives. Also, the ideas learned from both actual company experiences were shared with audiences in a new round of public speaking engagements.
Step 3: Write Journal Articles and Books:
The scholars write new papers and books that describe the enhanced theory and applications. For example, Cooper and Kaplan wrote about the extension of activity-based costing into measuring customer costing and profitability and, especially, the central role in activity-based costing for capacity costing and for distinguishing between costs of resources used vs. resources supplied (Cooper and Kaplan 1991a, 1991b, 1992). For the balanced scorecard, Norton and Kaplan wrote a second article (Kaplan & Norton 1993) and produced videotapes that featured the experiences and insights from Rockwater, FMC and Barclays Bank (Kaplan and Norton 1994).
After the second round of publications about activity-based costing and balanced scorecard, demand continued from companies for assistance in new implementations. By staying engaged with companies, the action researcher continues to extend the concepts further both to new industries (e.g., telecommunications, health care, government and nonprofits) and to new systems. Again (returning to Exhibit 1), the concept and practice of the innovation in leading-edge implementations was now at a significantly higher level than the early experiences reported at the beginning and end of the first loop around the innovation action research cycle.
Cooper and Kaplan believe that the next learning opportunity will come from incorporating activity-based costing concepts into the enterprise resource planning (ERP) packages that many companies are now installing.
For the balanced scorecard, the momentum from the early implementations, published articles, videos and speaking engagements, led to additional assignments with new companies. Norton and Kaplan navigated through their third loop around the innovation action research cycle. The balanced scorecard concept evolved extremely rapidly, even faster than activity-based costing. The findings from the third round of implementations were soon written up and published in three new cases (Kaplan 1995b; Kaplan 1996a; Kaplan 1996c); a third Harvard Business Review article (Kaplan and Norton 1996a); and a book (Kaplan and Norton 1996b). The balanced scorecard implementations being done at the end of 1995, as integrated strategic management systems, were far more advanced than the initial formulation, as a complementary nonfinancial measurement system.
Missing from the discussion so far is the role for evaluation. How can scholars determine whether a proposed new managerial concept, like activity-based costing or the balanced scorecard, is a good idea, one that creates value for organizations? Evaluation is important and scholars should be appropriately skeptical of claims associated with any new approach. The consulting and public speaking aspect of the innovation action research process makes many scholars even more skeptical and concerned. Now the researcher has become an advocate for the approach and can benefit financially through consulting relationships with companies and consulting firms, plus fees for speaking appearances. This intimate association with the concept and its implementation causes the action researcher to lose neutrality and be a less-than-credible source for independent evaluation.
In summary, implementation failures of even valid concepts can arise from three sources:
1) implementing the concept prematurely (without the oversight of the innovators) when the concept has yet to progress to be of most benefit to the organization;
2) poor management of the project that develops the new system; and
3) managers fail or refuse to act based on the signals from their new system. All of these sources make it difficult for passive, distant observers to perform credible evaluations of the new approach.
The problem of independence and objectivity for the action researcher remains. There is a tension between the intimate knowledge the action researcher has about the phenomenon, that would be highly valuable during evaluation studies, vs. the advocacy position of the action researcher to demonstrate success. One method for resolving this tension could be for the action researcher to be only one member of an evaluation team (or serve as a resource for such a team). The remaining scholars on the team would be independent, with their reputation, independence and objectivity serving to offset the perceived, and perhaps real, bias were action researchers to perform evaluation studies by themselves.
Another controversial aspect of the innovation action research is the compensation received by the action researchers for implementing the concept in organizations. Kaplan believes that compensation is critical for successful innovation action research. Companies should not view the involvement of action researchers with implementations (or their consulting alliance partners) as a favor they are doing for the researchers, granting them permission to do research on their sites. In Kaplan’s case, the compensation was a small, but important, hurdle to improve the likelihood that the project had strong support within the company.
See the following summaries for related Controversies
Argyris, C. and R. S. Kaplan. 1994. Implementing new knowledge: The case of activity-based costing. Accounting Horizons (September): 83-105. (Summary).
Norreklit, H. 2003. The balanced scorecard: What is the score? A rhetorical analysis of the balanced scorecard. Accounting, Organizations and Society 28(6): 591-619. (Summary).
Zimmerman, J. L. 2001. Conjectures regarding empirical managerial accounting research. Journal of Accounting and Economics (December): 411-427. (Summary).